Enzymatic synthesis of multiple spin-labeled DNA

Article abstract of DOI:10.1002/anie.200802314

(Chemical Equation Presented) Multi-spinning: The use of modified triphosphates and DNA polymerases has resulted in the multiple spin-labeling of DNA (see picture). A DNA helix entirely labeled with paramagnetic centers has been generated in this way and investigated by EPR spectroscopy.

Full text of DOI:10.1002/anie.200802314

Communications
DOI: 10.1002/anie.200802314
Spectroscopic Probes
Enzymatic Synthesis of Multiple Spin-Labeled DNA**
Samra Obeid, Maxim Yulikov, Gunnar Jeschke, and Andreas Marx*
Angewandte
Chemie
6782
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6782 –6785

Angewandte
Chemie
Electron paramagnetic resonance (EPR) spectroscopy is a
widespread technique for the study of the organizational and
dynamic properties of biological macromolecules. Many of
these applications depend on the sensitivity of nitroxide
labels[1] to dynamics on the picosecond to microsecond time
scales and on the ability to measure distances between such
labels on the nanometer length scale.[2] These techniques are
applicable in disordered systems, more sensitive than NMR
measurements, and provide more detailed information than
techniques based on optical excitation. Most biomacromole-
cules are diamagnetic in their native states and thus do not
have background EPR signals. Thus, spin-labeling techniques
can selectively address sites of interest in large molecules and
complex assemblies.[3] Attachment through a rigid linker is
necessary to minimize the effect of the motion of the spin
label on the spectrum and at the same time to maximize the
backbone dynamics of the biomacromolecule.[4] Such rigid
linkers are also favorable for distance measurements as they
lead to narrower distance distributions and thus to smaller
uncertainties in translating distances between labels to
structural models. However, the use of rigid linkers generates
the risk of perturbating the native structure because the label
cannot adapt to the steric requirements of its environment.
For this reason, labeling strategies have to be designed and
tested with great care.[5]
conjugated with suitably activated spin probes after their
chemical or enzymatic synthesis.[9] These methods allow the
incorporation of multiple probes into one molecule. However,
short linkers between the nucleic acid and the paramagnetic
center are still rather flexible, which have disadvantages for
the interpretation of the subsequent EPR spectra.
Here we present the site-specific introduction of spin
labels into DNA by using suitably modified nucleoside
triphosphates as building blocks in template-directed reac-
tions catalyzed by DNA polymerases. The synthesis of
multiple spin-labeled, long DNA oligonucleotides is feasible
by using this approach. We found that DNA polymerases of
eukaryotic, prokaryotic, and archaic origin can accommodate
the employment of spin-labeled nucleotides as building
blocks for DNA synthesis.
We first carried out the synthesis of the spin-labeled
nucleoside triphosphates 1 that contain nitroxide-based para-
magnetic centers connected to the nucleobase (Scheme 1).
Recently, EPR spectroscopy was applied extensively in
studies on the structures and dynamics of nucleic acids.[1,6]
Since nucleic acids do not contain any natural paramagnetic
centers, spin labels have to be introduced prior to EPR
investigations. Several methods have previously been estab-
lished for the introduction of a paramagnetic center, for
example, a stable nitroxide, at a specific site in DNA. Such
spin labels were introduced either by employment of a spin-
labeled building block during automatic DNA synthesis,[7] or
functionalized building blocks were introduced into the
growing DNA first and subsequently coupled to a spin label
on a solid support, for example, by employment of palladium-
catalyzed coupling reactions.[8] Single-labeled oligonucleoti-
des with relatively short lengths have been synthesized by
these techniques. However, the length of the oligonucleotides
and their degree of modification is restricted by the inherent
limitations of automatic DNA synthesis. To the best of our
knowledge, multiple site-specific incorporation of spin labels
into DNA by using these methods has not as yet been
demonstrated. Other approaches are based on the incorpo-
ration of additional functionalities in nucleic acids that are
Scheme 1. Synthesis of spin-labeled TTP analogues 1. Conditions and
reagents: a) (2-(4-iodophenyl)ethynyl)trimethylsilane, [Pd(PPh3)4], CuI,
DMF, microwave, 15 min, 1008C, 64%; b) tetrabutylammonium fluo-
ride, CH2Cl2, 5 min, 08C, 83%; c) H-2, [Pd(PPh3)4], CuI, DMF, 43%
(5a); d) H-3, [Pd(PPh3)4], CuI, DMF, 67% (5b); e) 2-chloro-4H-1,3,2-
benzodioxaphosphorin-4-one, (nBu3NH)2H2P2O7, I2, pyridine, H2O,
then NH3, 31% (1a); f) 2-chloro-4H-1,3,2-benzodioxaphosphorin-4-
one, (nBu3NH)2H2P2O7, then I2, pyridine, H2O, then NH3, 31% (1b).
We chose modification at the C5-position in 2’-deoxyuridine
since modifications at this position do not significantly
interfere with Watson–Crick base pairing. Furthermore, it
has been shown that DNA polymerases are able to tolerate
modifications at this position at least to a certain extent, as
seen in the substitution of thymidine with other base
analogues.[10,11] In addition, the palladium-catalyzed cross-cou-
pling of alkynes to the respective 5-iodo-2’-deoxyuridine
analogues has been demonstrated.[12,13] To assess the effect of
the length of the linker that connects the spin label with the
nucleotide on the ability of DNA polymerases to accept 1 we
synthesized alkyne-modified nitroxide H-3 starting from known
H-2[7b,14] by a Sonogashira cross-coupling reaction. Nitroxides H-
2 and H-3 were coupled with the protected 5-iodo-2’-deoxyur-
idine derivative 4 to yield 5, which was subsequently trans-
formed into the nucleoside triphosphates 1.[15]
[*] S. Obeid, Dr. M. Yulikov,[+] Prof. Dr. G. Jeschke,[+] Prof. Dr. A. Marx
Department of Chemistry and
Konstanz Research School Chemical Biology
University of Konstanz
Universitätsstrasse 10, 78457 Konstanz (Germany)
Fax: (+49) 7531-88-5140
E-mail: andreas.marx@uni-konstanz.de
[+] Present address:
Department of Chemistry, ETH Zurich
8093 Zurich (Switzerland).
[**] We gratefully acknowledge funding by the DFG.
We next studied the action of nucleoside triphosphates 1
on DNA polymerases by investigating the propensity of the
Hintergrundinformationen zu diesem Beitrag sind im WWW unter
http://dx.doi.org/10.1002/ange.200802314 zufinden.
Angew. Chem. Int. Ed. 2008, 47, 6782 –6785
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6783

Communications
Klenow fragment of E. coli DNA polymerase I (3’!5’-
(Figure 1d, traces I and II) can be fitted well by assuming an
isotropic rotational diffusion in the fast regime,[16] with
rotational correlation times tc of 57 and 95 ps, respectively
(see the Supporting Information). The increase in the value of
tc with increasing linker length is in line with expectations.
After the incorporation (traces III and IV), the model of
isotropic rotational diffusion no longer fits the spectra. The
relative line intensities are consistent with a preferred motion
about the linker backbone axis. Molecular rotation perpen-
dicular to this axis is slower, but is still well within the fast
regime for the shorter linker (trace III). This result indicates
rotation about the DNA helix axis on a sub-nanosecond time
scale. This rotation is slowed down for the longer linker
(trace IV) and approaches the nanosecond time scale. As the
modeling of the spectral line shapes are ambiguous for
anisotropic rotational diffusion in the fast regime, we refrain
from deriving rotational diffusion tensors.
exonuclease deficient variant, KF(exo-)) and human DNA
polymerase b (Polb) to accept 1 in primer extension experi-
ments. First we used a 5’-end 32P-labeled 23-nucleotide primer
and 35-nucleotide template complex that leads to incorpo-
ration of a thymidine analogue at position 27 after incorpo-
ration of three nucleotides (Figure 1a). We found for both
Encouraged by these initial findings, we next investigated
whether multiple spin labels could be incorporated within one
helix by DNA polymerases. For this we designed two
templates in a way that they code for the incorporation of a
spin-labeled nucleotide at every second or every fourth
nucleotide. A third template was designed so that 11 adjacent
2’-deoxyadenosine residues code for the incorporation of 11
spin-labeled nucleotides in a row and form one entire DNA
helix turn of spin labels. Preliminary studies of several DNA
polymerases indicated that commercially available Thermi-
nator DNA polymerase (A485L mutant of Thermococcus
species 98N DNA polymerase) is most proficient in processing
multiple spin-labeled nucleotides 1a and 1b. Furthermore, it
turned out that this enzyme accepted 1a better than 1b. Thus,
for further studies we used 1a exclusively. Figure 2 shows the
results obtained in multiple incorporation experiments by
using Therminator DNA polymerase. Full-length products
were obtained in all cases when natural TTP was substituted
by 1a. Again, slower migrating products were observed by
PAGE analysis when the bulkier 1a was used instead of
natural TTP.
Interestingly, Therminator DNA polymerase is proficient
in incorporating 11 adjacent modified nucleotides of 1a,
thereby resulting in one entire DNA helical turn decorated
with multiple spin systems (Figure 2c). The reaction product
was purified by HPLC and further characterized by thermal
denaturation (Tm) measurements as well as by CD and EPR
spectroscopy. The CD spectrum indicates that the intensively
modified DNA bearing 11 adjacent spin probes indeed adopts
an overall B-form conformation similar to the unmodified
DNA (see the Supporting Information). In addition, the
Tm measurements indicate only a small effect of the modifi-
cations on the stability of the duplex (see the Supporting
Information).
Figure 1. Incorporation of 1a,b and EPR investigations. a) Partial
sequences of the primer and template. b) Primer extension studies
employing KF(exo-): lane 0: primer only, lane 1: primer extension
performed in the presence of dATP, dGTP, and dCTP, but without any
TTP analogue; lane 2: as lane 1, but in the presence of TTP; lane 3: as
lane 1, but in the presence of 1a; lane 4: as lane 1, but in the presence
of 1b. c) As in (b), but with the use of Polb instead of KF(exo-).
d) EPR spectra of 1a (I), 1b (II), the product derived from the reaction
in (b), lane 3, by using 1a (III), and the product derived from the
reaction in (b), lane 4, by using 1b (IV). Further experimental details
are provided in the Supporting Information.
enzymes that primer elongation is paused at positions 26 or 27
when the reactions were performed in the presence of only
dATP, dGTP, and dCTP (Figure 1b, c; lanes 1). When all four
natural dNTPs were added, both enzymes resulted in the
formation of the full-length product (Figure 1b, c; lanes 2).
When the natural thymidine triphosphate (TTP) was sub-
stituted by 1a or 1b, respectively, reaction products were
formed that migrated slower in denaturing polyacrylamide gel
electrophoresis (PAGE) than the original band corresponding
to the full-length products derived from experiments with
TTP (Figure 1b, c; lanes 3 and 4). The retardation can be
explained by the additional bulk of the modifications. Similar
effects have been reported before.[10] Interestingly, significant
pausing bands were observed at position 27 when 1a was
employed in the KF(exo-)-catalyzed reaction (Figure 1b,
lane3). This finding indicates that further extension is some-
how hampered after incorporation of 1a. No such bands were
observed when nucleotide 1b was employed, which indicates
that the bulky nitroxide group is better tolerated by this
enzyme when placed on an extended linker. After purification
of the enzymatically synthesized DNA by HPLC, the CD
spectra indicates that the modified oligonucleotides adopt a
similar overall B-form conformation as the unmodified
strands (see the Supporting Information).
Surprisingly, the EPR spectrum (Figure 3) indicates a
reduced anisotropy of rotational diffusion compared to the
singly labeled DNA. However, unlike in the spectrum of 1a,
the low-field line still has a significantly smaller amplitude
than the center line, which demonstrates that the motion is
not completely isotropic. The absence of line-shape effects
arising from through-space exchange coupling, except for the
additional homogeneous line broadening compared to the
The successful incorporation of the nucleotides derived
from 1a and 1b was further verified by ESI mass spectrom-
etry and EPR spectroscopy. The EPR spectra of 1a and 1b
6784 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6782 –6785

Angewandte
Chemie
proficiency to be accepted as substrates by DNA polymerases.
DNA polymerases of eukaryotic, prokaryotic, and archaic
origin are competent in the employment of spin-labeled
nucleotides as surrogates of natural building blocks in enzy-
matic DNA synthesis. This finding opens up new opportunities
for further applications, for example, for in vivo spin labeling or
the generation of complex multiple spin systems, which might
be useful for DNA-based nanobiotechnology. Further inves-
tigations along these lines are in progress.
Received: May 17, 2008
Published online: July 31, 2008
Keywords: DNA polymerase · DNA · EPR spectroscopy ·
.
nucleotides · spin labels
[1] Spin Labeling: The Next Millennium (Biol. Magn. Reson. 1998, 14).
[2] a) G. Jeschke, Y. Polyhach, Phys. Chem. Chem. Phys. 2007, 9,
1895; b) O. Schiemann, T. F. Prisner, Q. Rev. Biophys. 2007 40, 1.
[3] W. L. Hubbell, D. S. Cafiso, C. Altenbach, Nat. Struct. Biol. 2000,
7, 735.
[4] A. Spaltenstein, B. H. Robinson, P. B. Hopkins, J. Am. Chem.
Soc. 1988, 110, 1299.
[5] O. Schiemann, N. Piton, J. Plackmeyer, B. E. Bode, T. F. Prisner,
J. W. Engels, Nat. Protocols 2007, 2, 904.
[6] a) P. Z. Qin, I. S. Haworth, Q. Cai, A. K. Kusnetzow, G. P. G.
Grant, E. A. Price, G. Z. Sowa, A. Popova, B. Herreros, H. He,
Nat. Protocols 2007, 2, 2354; b) R. Ward, D. J. Keeble, H. El-
Mkami, D. G. Norman, ChemBioChem 2007, 8, 1957; c) N.
Barhate, P. Cekan, A. P. Massey, S. T. Sigurdsson, Angew. Chem.
2007, 119, 2709; Angew. Chem. Int. Ed. 2007, 46, 2655.
[7] a) A. Spaltenstein, B. H. Robinson, P. B. Hopkins, J. Am. Chem.
Soc. 1988, 110, 1299; b) A. Spaltenstein, B. H. Robinson, P. B.
Hopkins, Biochemistry 1989, 28, 9484.
Figure 2. Multiple spin labeling under Therminator DNA polymerase
catalysis and incorporation of 1a. The primer template sequences
employed are indicated above each panel. Lanes 0: primer only;
lanes 1: primer extension performed in the presence of dATP, dGTP,
and dCTP, but without any TTP analogue; lanes 2: as in lanes 1, but in
the presence of TTP; lanes 3: as in lanes 1, but in the presence of 1a.
[8] a) O. Schiemann, N. Piton, Y. Mu, G. Stock, J. W. Engels, T. F.
Prisner, J. Am. Chem. Soc. 2004, 126, 5722; b) O. Schiemann, N.
Piton, J. Plackmeyer, B. E. Bode, T. F. Prisner, J. W. Engels,
Nucleosides Nucleotides Nucleic Acids 2005, 24, 771; c) N. Piton,
Y. Mu, G. Stock, T. F. Prisner, O. Schiemann, J. W. Engels,
Nucleic Acids Res. 2007, 35, 3128.
[9] a) P. W. Langenmeier, A. M. Bobst, Arch. Biochem. Biophys.
1981, 208, 205; b) A. M. Bobst, S.-C. Kao, R. C. Toppin, J. C.
Ireland, I. E. Thomas, J. Mol. Biol. 1984, 173, 63.
Figure 3. EPR spectra of a 49nt DNA duplex containing 11 adjacent
spin labeled nucleotides derived from 1a.
[10] a) O. Thum, S. Jäger, M. Famulok, Angew. Chem. 2001, 113,
4112; Angew. Chem. Int. Ed. 2001, 40, 3990; b) S. Jäger, M.
Famulok, Angew. Chem. 2004, 116, 3399; Angew. Chem. Int. Ed.
2004, 43, 3337; c) S. Jäger, G. Rasched, H. Kornreich-Leshem, M.
Engeser, O. Thum, M. Famulok, J. Am. Chem. Soc. 2005, 127,
spectra of 1a, is feasible for a rigid structure. Solid-state
spectra that would have been conclusive about the presence
or absence of dipole–dipole coupling between the spins could
not be obtained with the currently available sample concen-
tration. We do find, however, that the intensity of the
spectrum is consistent with the presence of 11 spin labels
per DNA molecule at the given DNA concentration. The
EPR spectra are thus also in agreement with the proposed
structure, although it is unclear at present why the motion
about the helix axis is faster for the multiply labeled 49mer
than for the singly labeled 35mer.
ˇ
15071; d) P. Capek, H. Cahovꢀ, R. Pohl, M. Hocek, C.
Gloeckner, A. Marx, Chem. Eur. J. 2007, 13, 6115.
[11] H. Sawai, A. N. Ozaki, F. Satoh, T. Ohbayashi, M. M. Masud, H.
Osaki, Chem. Commun. 2001, 2604.
[12] a) M. J. Robbins, P. Barr, Tetrahedron Lett. 1981, 22, 421;
b) F. W. J. Hobbs, J. Org. Chem. 1989, 54, 3420; c) J. D. Kahl,
M. M. Greenberg, J. Am. Chem. Soc. 1999, 121, 597.
[13] a) L. A. Agrofoglio, I. Gillaizeau, Y. Saito, Chem. Rev. 2003, 103,
1875; b) R. Chinhilla, C. Najera, Chem. Rev. 2007, 107, 874.
[14] S. W. Stork, M. W. Makinen, Synthesis 1999, 1309.
[15] J. Ludwig, F. Eckstein, J. Org. Chem. 1989, 54, 631.
[16] J. H. Freed, G. K. Fraenkel, J. Chem. Phys. 1963, 39, 326.
In summary, we have demonstrated for the first time the
synthesis of spin-labeled nucleoside triphosphates 1 and their
Angew. Chem. Int. Ed. 2008, 47, 6782 –6785
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6785